Posted By
SukkoPera
on 2022-08-26
10:56:50
|  Re: Solder's SIDcard Replica
Oh, I see  . The 245 is only used for reading the digital joystick buttons though, it wouldn't hurt the audio part even if I hadn't included it (and in fact I haven't mounted it on one of my test boards).
I think I have found the culprit! The 4520 used on Luca's board was produced by RFT electronic (in East Germany?), here's its datasheet:
Notice the propagation time (Verzoegerungszeit): 560 ns (max)! The 4520 used on Chronos's board was made by SGS-Thomson instead but it has exactly the same figure. This one also lists the typical propagation time, which is 280 ns.
These times are *huge*, basically an order of magnitude greater than the modern chip I'm using (55 ns max!). Those chips are so slow that the SID PHI2 they generate will be shifted significantly, so much that it will probably be still high for some time after the second time SID /CS is asserted in double-clock periods, allowing the write to get through.
I have just managed to track down some of those older chips, but since it will take a while for them to get here, I have something in mind to test this theory with the chips I'm currently using. If it works, I'll integrate the solution in the next version of the board, so that one doesn't have to track down NOS to build it .
We're almost there, stay tuned!
|
